Unsupervised Learning of Noisy-Or Bayesian Networks
This provides a more efficient approach for learning large Bayesian networks in medical diagnosis, though it is incremental as it builds on prior work for discrete-valued mixture models.
The paper tackles the problem of learning parameters in Bayesian networks with hidden variables by introducing a polynomial-time method-of-moments algorithm that avoids inference during learning, and demonstrates its application to fully learn the parameters of the QMR-DT medical diagnosis network with only findings observed.
This paper considers the problem of learning the parameters in Bayesian networks of discrete variables with known structure and hidden variables. Previous approaches in these settings typically use expectation maximization; when the network has high treewidth, the required expectations might be approximated using Monte Carlo or variational methods. We show how to avoid inference altogether during learning by giving a polynomial-time algorithm based on the method-of-moments, building upon recent work on learning discrete-valued mixture models. In particular, we show how to learn the parameters for a family of bipartite noisy-or Bayesian networks. In our experimental results, we demonstrate an application of our algorithm to learning QMR-DT, a large Bayesian network used for medical diagnosis. We show that it is possible to fully learn the parameters of QMR-DT even when only the findings are observed in the training data (ground truth diseases unknown).